8.10
Contingency Table
A contingency table provides a way of portraying data that can facilitate calculating probabilities. It is a method of displaying a frequency distribu…
The contingency table, cross-tabulation, or two-way frequency table, statistically summarizes a relationship between two categorical variables.
Here, the outcome of one variable is arranged in rows, while that of the other variable is arranged in columns.
Consider two variables: COVID-19 cases and vaccination status. The contingency table for this data will contain two rows and three columns. The row variable identifies whether individuals are infected or not. The column variable identifies the individual's vaccination status.
Of the 236 total observations, 150 individuals are infected, and 86 are not. Furthermore, 129 individuals are unvaccinated, 27 are partially vaccinated, and 80 are completely vaccinated.
The contingency table displays the observed frequencies of two variables. After performing a test of independence, the chi-square statistic is used to determine whether these values mirror the expected frequency.
Contingency tables help to display and find the probability of any event occurring dependent or independent of other variables, which are visualized using histograms or bar graphs. Thus, they help systematically organize, analyze and quantify data.
View the full transcript and gain access to JoVE Core videos
Q1: What is a contingency table and how is it structured?
A contingency table, also called a cross-tabulation or two-way frequency table, summarizes the relationship between two categorical variables. One variable's outcomes are arranged in rows, while the other variable's outcomes are arranged in columns. This structure displays observed frequencies and helps systematically organize and analyze data from two variables.
Q2: How do you read a contingency table with COVID-19 and vaccination data?
In a COVID-19 and vaccination contingency table, rows indicate infection status (infected or not infected), while columns show vaccination status (unvaccinated, partially vaccinated, or completely vaccinated). The table displays how many individuals fall into each combination. For example, of 236 total observations, 150 were infected and 86 were not infected across all vaccination categories.
Q3: What role does the chi-square statistic play in contingency tables?
After performing a test of independence, the chi-square statistic determines whether observed frequencies in a contingency table match expected frequencies. This comparison helps establish whether two categorical variables are independent or dependent on each other, revealing patterns in the data.
Q4: How do contingency tables help calculate probabilities?
Contingency tables display observed frequencies of two variables, enabling quick calculation of conditional probabilities. By organizing data into rows and columns, they show how one variable may depend on another, allowing students to determine the probability of events occurring dependent or independent of other variables.
Q5: What is the difference between observed and expected frequencies in contingency tables?
Observed frequencies are the actual counts displayed in a contingency table showing how many cases fall into each category combination. Expected frequencies are theoretical values calculated under the assumption of independence. A test of independence compares these two to determine whether variables are truly dependent or independent.
Q6: How can contingency tables be visualized in statistical analysis?
Contingency tables help display and find probabilities of events, which are often visualized using histograms or bar graphs. These visual representations make it easier to interpret relationships between categorical variables and communicate findings about data patterns and dependencies to audiences.
Q7: Why are contingency tables important for organizing categorical data?
Contingency tables systematically organize, analyze, and quantify data by displaying how two categorical variables may be dependent or contingent upon each other. They provide a structured method for portraying frequency distributions that facilitates calculating probabilities and understanding relationships between variables in research.